The route to extinction: population dynamics of a threatened butterfly.

We compare results of field study and model analysis of two butterfly populations to evaluate the importance of alternative mechanisms causing changes in abundance. Although understanding and predicting population fluctuations is a central goal of population ecology, it is not often achieved because long-term abundance data are available for few populations in which mechanisms causing fluctuations also are known. Both kinds of information exist for two populations of the checkerspot butterfly, Euphydryas editha bayensis, which are matched in most ways except for habitat area and topography. We applied results from field study to make predictions about the dynamics of the two populations. Then we tested these predictions using nonlinear modeling of abundance data. Models included endogenous factors, exogenous effects of weather, or both. Results showed that the populations differed in variability and responses to endogenous and exogenous factors. The population in the more homogeneous habitat varied more widely, went extinct first, and fluctuated more severely with climate. Dynamics of the population occupying the topographically diverse habitat were more complex, containing damped oscillations and weaker influences of weather. We draw four main conclusions. First, the routes to extinction for E. e. bayensis populations in protected habitat were random walks driven by climatic variability. Climatic influences dominated both populations, but the timing and functional forms of climatic effects differed between populations. Second, topographic diversity reduced weather-induced population variability and increased persistence time. Third, one must explicitly consider both endogenous and exogenous components to fully understand population dynamics. Fourth, resolving the debate over population regulation requires integrating long-term population sampling, model analysis, and investigation of mechanisms in the field.